Original Article
Clinical, radiological, and histopathological features of checkpoint inhibitor-related pneumonitis: a retrospective cross-sectional analysis of 44 patients who underwent lung biopsy
Abstract
Background: Checkpoint inhibitor-related pneumonitis (CIP) is a serious immune-related adverse event associated with relatively high mortality. Its clinical and radiological manifestations are non-specific and may overlap with infection, tumour progression, and other pulmonary diseases. Currently, the histopathological spectrum of CIP and its relationship with radiological patterns and clinical outcomes remain insufficiently defined. This study aimed to analyse the clinical, radiological, and histopathological features, as well as exploratory prognostic outcomes, of biopsy-evaluated patients with clinically diagnosed or highly suspected CIP.
Methods: We retrospectively reviewed patients who were clinically diagnosed with or highly suspected of having CIP and underwent lung biopsy at The First Affiliated Hospital of Guangzhou Medical University between June 2018 and June 2022. Medical records were reviewed to collect clinical characteristics, radiological findings, laboratory parameters, pathological results, treatment information, and follow-up data. Chest images were independently assessed by two thoracic radiologists, and pathological slides were interpreted by pulmonary pathologists. Survival and prognostic data were obtained from medical records and follow-up information.
Results: A total of 217 patients clinically diagnosed with or highly suspected of having CIP were identified, of whom 45 (20.7%) underwent lung biopsy. One patient was excluded because adequate lung tissue was not obtained, leaving 44 patients for the final analysis. Seven patients underwent ultrasonography-guided percutaneous transthoracic needle biopsy (USG-TTNB), and 37 underwent transbronchial lung biopsy (TBLB); biopsy-related complications included two cases of pneumothorax and one case of haemorrhage. The main radiological patterns were organising pneumonia (OP) and nonspecific interstitial pneumonia (NSIP). Histopathological patterns included isolated NSIP, NSIP with OP (NSIP + OP), OP, usual interstitial pneumonia (UIP), and mild interstitial pneumonitis (mild IP). The concordance rate between radiological patterns and histopathological classifications was only 29.5%, with an exploratory Cohen’s kappa value of 0.04. Exploratory survival analysis showed no significant difference in overall survival among patients with OP, NSIP, and NSIP + OP histopathology (P=0.71).
Conclusions: For patients with clinically suspected but diagnostically challenging CIP, TBLB or USG-TTNB selected after multidisciplinary assessment may provide histopathological information and help exclude alternative diagnoses such as infection, tumour progression, and other lung injuries. The marked heterogeneity of histopathological findings and the low radiology–pathology concordance suggest that CIP assessment should not rely solely on either radiological or pathological findings. Multidisciplinary assessment may improve diagnostic confidence and support individualised treatment decisions, whereas the prognostic significance of different histopathological patterns requires validation in larger studies.

